多传感器数据融合系统中联合概率数据互联算法的研究

Study of multi-sensor joint probabilistic data association in multi-sensor data fusion system

概述了多传感器数据融合系统中的联合概率数据互联算法,给出了MSJPDA的两种处理结构,分析了其算法的复杂度。并在此基础上,结合B.Zhou提出的直接概率计算和近似概率计算的方法,提出了一种基于近似聚的近似概率数据互联算法(MSJPDA),通过仿真研究以及和最近邻法所做的比较表明,该方法确实能提高在密集情况下的数据融合精度,算法耗时与最近邻法相差不大,精确度接近完全概率互联算法。

Multi-ensor Joint Probabilistic Data Association(MSJPDA) was presented. Approximate Multi-sensor Joint Probabilistic Data Association(AMSJPDA) as a new methodology was proposed by combining B. Zhou's theory of the approximate probabilistic computing and direct probabilistic computing. The comparing AMSJPDA with NN (Nearest Neighbor) showed that using AMSJPDA could improve the precision of association in a complex environment. It demanded only a little more time than NN and the precision was as good as MSJPDA.